public static class BalancedShardsAllocator.WeightFunction extends Object
index balance
- balance property over shards per indexshard balance
- balance property over shards per clusterEach of these properties are expressed as factor such that the properties factor defines the relative importance of the property for the weight function. For example if the weight function should calculate the weights only based on a global (shard) balance the index balance can be set to 0.0 and will in turn have no effect on the distribution.
The weight per index is calculated based on the following formula:weightindex(node, index) = indexBalance * (node.numShards(index) - avgShardsPerNode(index))
weightnode(node, index) = shardBalance * (node.numShards() - avgShardsPerNode)
weight(node, index) = weightindex(node, index) + weightnode(node, index)
Constructor and Description |
---|
WeightFunction(float indexBalance,
float shardBalance) |
Modifier and Type | Method and Description |
---|---|
float |
weight(BalancedShardsAllocator.Balancer balancer,
org.elasticsearch.cluster.routing.allocation.allocator.BalancedShardsAllocator.ModelNode node,
String index) |
float |
weightShardAdded(BalancedShardsAllocator.Balancer balancer,
org.elasticsearch.cluster.routing.allocation.allocator.BalancedShardsAllocator.ModelNode node,
String index) |
float |
weightShardRemoved(BalancedShardsAllocator.Balancer balancer,
org.elasticsearch.cluster.routing.allocation.allocator.BalancedShardsAllocator.ModelNode node,
String index) |
public float weight(BalancedShardsAllocator.Balancer balancer, org.elasticsearch.cluster.routing.allocation.allocator.BalancedShardsAllocator.ModelNode node, String index)
public float weightShardAdded(BalancedShardsAllocator.Balancer balancer, org.elasticsearch.cluster.routing.allocation.allocator.BalancedShardsAllocator.ModelNode node, String index)
public float weightShardRemoved(BalancedShardsAllocator.Balancer balancer, org.elasticsearch.cluster.routing.allocation.allocator.BalancedShardsAllocator.ModelNode node, String index)
Copyright © 2009–2016. All rights reserved.